Survey

Data preparation

This is an R Markdown document. Markdown is a simple formatting syntax for authoring HTML, PDF, and MS Word documents. For more details on using R Markdown see http://rmarkdown.rstudio.com.

lang <- params$lng

library(tidyverse)
library(hrbrthemes)
#remotes::install_github("hadley/emo")
library(emo)

dt <- read_csv("data/WintonCentreCleaned_covid_8country_labelled.csv") 

# Make a small df with codes corresponding to missing values
msng_df <- tibble::tribble(~var, ~code, ~msng,
                           "Ethnic min", "4",  TRUE,
                           "Govrestrict_1","8", TRUE,
                           "Govrestrict_2","8", TRUE,
                           "Govrestrict_3","8", TRUE )

country_df <- tibble::tribble( ~Residency, ~ji_name, ~countryname_en, ~countryname_se,
                               "AU", "australia", "Australia", "Australien",
                               "DE", "de", "Germany", "Tyskland",
                               "ES", "es", "Spain", "Spanien",
                               "GB", "gb", "UK", "Storbritannien",
                               "IT", "it", "Italy", "Italien",
                               "MX", "mexico","Mexico", "Mexico",
                               "SE", "sweden", "Sweden", "Sverige",
                               "US", "us", "USA", "USA" )

country_df <- country_df %>% 
  pivot_longer(starts_with("countryname_"), names_to = "language", 
               names_prefix = "countryname_", values_to = "countryname") %>% 
  filter(language==lang) %>% 
  select(-language) %>% 
  mutate(countryflag=map_chr(ji_name, emo::ji))

wrap_txt <- function(x, w=80){
  paste(strwrap(x,
    width = w
  ),
  collapse = "\n")
}

# Prepare labels for annotating plots
q_df_raw <- dt %>% slice(1:2) %>% 
  tibble::rowid_to_column(var = "row_id") %>% 
  pivot_longer(-row_id, names_to = "var", values_to = "txt_en") %>% 
  arrange(var, row_id)

write_csv(q_df_raw, "data/q_df_raw_en.csv")
q_df <- read_csv("data/q_df_en_se.csv")

q_df <- q_df %>% 
  pivot_longer(starts_with("txt"), names_to = "language", 
               names_prefix = "txt_", values_to = "question") %>% 
  filter(language==lang)%>% 
  select(-language) %>% 
  replace_na(list(question=""))

lbl_df_raw <- dt %>% slice(3) %>% 
  map(str_replace, " to (?=\\d)| - (?=\\d)", ", ") %>% 
  map(str_replace_all, "(?<=\\d\\s)\\(", " = ") %>% 
  map_df(~enframe(str_split(.x, ",\\s+(?=\\d)")[[1]] ), .id = "var") %>%
  select(-name) %>% 
  separate(value, sep="\\s?+=\\s?+", into=c("code", "label"), extra="merge", convert=TRUE) %>% 
  mutate(label=str_replace(label, "\\)$", ""),
         label=str_replace_all(label, "'", ""),
         label=trimws(label)) %>% 
  mutate(code=as.character(code))

write_csv(lbl_df_raw, "data/lbl_df_raw_en.csv")
lbl_df <- read_csv("data/lbl_df_en_sv.csv")

lbl_df <- lbl_df %>% 
  pivot_longer(starts_with("label"), names_to = "language", 
               names_prefix = "label_", values_to = "label") %>% 
  filter(language==lang)%>% 
  select(-language) %>% 
  anti_join(type_convert(msng_df), by=c("var", "code"))

# Prepare data
df_lng_raw <- dt %>% 
  slice(-1:-3) %>% 
  mutate(Residency=ifelse(Residency=="UK", "GB", Residency)) %>% 
  tibble::rowid_to_column() %>% 
  pivot_longer(cols = GenSocTrust:Politics, names_to = "var", values_to = "code") %>% 
  left_join(msng_df, by=c("var", "code")) %>% 
  replace_na(list(msng=FALSE))

write_csv(df_lng_raw, "data/df_long_raw.csv")

data_df <- df_lng_raw %>% 
  mutate(code=ifelse(msng, NA, code),
         code_num = parse_number(code)) %>% 
  count(var, Residency, code, code_num) %>% 
  group_by(var) %>% 
  mutate(mid_category_num = ceiling(mean(code_num, na.rm = TRUE))) %>% 
  group_by(var, Residency) %>% 
  mutate(nna_n=ifelse(!is.na(code), n, NA),
         pct=n/sum(n),
         nna_pct = nna_n/sum(nna_n, na.rm=TRUE),
         sgnd_pct=ifelse(code_num >= mid_category_num, pct, -pct),
         nna_sgnd_pct=ifelse(code_num >= mid_category_num, nna_pct, -nna_pct),
         sum_pct_low = -sum(pct * (code_num < mid_category_num), na.rm = TRUE),
         sum_pct_hi = sum(pct * (code_num > mid_category_num), na.rm = TRUE),
         panel=case_when(
             is.na(code_num) ~ "Missing",
             code_num == mid_category_num ~ "Neutral",
             TRUE ~ "Responses") 
           ) %>% 
      ungroup() %>% 
    left_join(country_df, by="Residency")

# Service function for single questions
plot_single <- function(var_name, q_dfa=q_df, lbl_dfa=lbl_df, data_dfa=data_df){
 lbl_var_df <- lbl_dfa %>%
     filter(var==var_name) %>%
     arrange(desc(code)) %>% # desc(code)
     #mutate(x01=(1:n())-1)
     mutate(x01=1-(code-min(code))/(max(code)-min(code)))

 q_var_df <- q_dfa %>% 
  filter(var==var_name)

 clrs <- RColorBrewer::brewer.pal(max(lbl_var_df$code), "BrBG")
 lbl_var_df$clr <- clrs[as.numeric(lbl_var_df$code)]
  
 df_plot <- data_dfa %>% 
    filter(var==var_name)
 
 df_plot %>%
    filter(!is.na(code)) %>% 
    ggplot(mapping=aes(y=fct_rev(countryflag)))+
    geom_col(mapping=aes(x=pct, fill=(code)),
               position = "fill") +
    geom_text(data=lbl_var_df,
               aes(x=x01, y=Inf, color=I(clr), label=label),
              family="Roboto Condensed",
               hjust="inward", vjust=1) +
    scale_fill_manual(values=clrs ) +
    theme_ipsum_rc(grid=FALSE)+
    scale_y_discrete(expand =  expansion(add = c(0,1)))+
    scale_x_continuous(labels = scales::percent, expand = expansion(add=c(0.01, 0.02)),
                         sec.axis = sec_axis( ~1-., labels = scales::percent))+
    labs(title=wrap_txt(with(q_var_df, question[row_id==1]), 75), 
         #subtitle = with(q_var_df, question[row_id==2]), 
         y=NULL, x=NULL) + 
    theme(panel.grid.major = element_blank(),
          panel.grid.minor = element_blank(),
          axis.line.y = element_blank(),
          axis.ticks.y = element_blank(),
          axis.text.y = element_text(size=16),
          plot.title = element_text(lineheight = 1.1, face="bold", size=14),
          legend.position = 'none')
}

# Service function for group questions
plot_group <- function(var_group_name, q_dfa=q_df, lbl_dfa=lbl_df, data_dfa=data_df) {
 q_var_df <- q_dfa %>% 
  filter(var_group==var_group_name, row_id==2) %>% 
  select(var, question)

 q_title <- q_dfa %>% 
  filter(var_group==var_group_name, row_id==0) %>% 
  pull(question)

 var_names_df <- distinct(q_var_df, var)
 
 lbl_var_df <- lbl_dfa %>%
     semi_join(var_names_df, by="var") %>% 
     group_by(code) %>% 
     summarise(label=first(label)) %>% 
     arrange(desc(code)) %>% 
     mutate(x01=1-(code-min(code))/(max(code)-min(code)))

 clrs <- RColorBrewer::brewer.pal(max(lbl_var_df$code), "BrBG")
 lbl_var_df$clr <- clrs[as.numeric(lbl_var_df$code)]
  

 df_plot <- data_dfa %>% 
    semi_join(var_names_df, by="var") %>% 
    left_join(q_var_df, by="var")

 df_plot %>%
    filter(!is.na(code)) %>% 
    mutate(question=map_chr(question, wrap_txt, 75)) %>% 
    ggplot(mapping=aes(y=fct_rev(countryflag)))+
    geom_col(mapping=aes(x=pct, fill=code),
               position = "fill") +
    geom_text(data=lbl_var_df,
               aes(x=x01, y=0.4, color=I(clr), label=label),
               family="Roboto Condensed",
               hjust="inward", vjust=1) +
    facet_wrap(vars(question), ncol = 2)+
    scale_fill_manual(values=clrs) +
    theme_ipsum_rc(grid=FALSE)+
    scale_y_discrete(expand =  expansion(add = c(1,0)))+
    scale_x_continuous(labels = scales::percent, expand = expansion(add=c(0.01, 0.02)),
                         sec.axis = sec_axis( ~1-., labels = scales::percent))+
    labs(title=wrap_txt(q_title,105), 
         y=NULL, x=NULL) + 
    theme(panel.grid.minor = element_blank(),
          panel.grid.major = element_blank(),
          axis.line.y = element_blank(),
          axis.ticks = element_blank(),
          axis.text.y = element_text(size=20),
          plot.title = element_text(lineheight = 1.1, face="bold"),
          strip.text = element_text(size=12),
          legend.position = 'none')
}

# Service function for categorical plots by country
plot_categorical <- function(var_name, q_dfa=q_df, lbl_dfa=lbl_df, data_dfa=data_df){

 lbl_var_df <- lbl_dfa %>%
     filter(var==var_name) %>%
     arrange(desc(code)) %>% 
     select(-var) %>% 
     mutate(code=as.character(code))

 q_var_df <- q_dfa %>% 
  filter(var==var_name)

 clrs <- RColorBrewer::brewer.pal(7, "BrBG")[7]
 lbl_var_df$clr <- clrs[as.numeric(lbl_var_df$code)]
  

 df_plot <- data_dfa %>% 
    filter(var==var_name)
 
 df_plot %>% 
    filter(!is.na(code)) %>%
    left_join(lbl_var_df, by="code") %>% 
    ggplot(mapping=aes(y=fct_rev(label)))+
    geom_col(mapping=aes(x=pct), fill=clrs, show.legend = FALSE)+
    facet_wrap(vars(paste(countryflag, countryname)), ncol = 2)+
    theme_ipsum_rc(grid = FALSE)+
   # scale_fill_manual(values=clrs)+
   #scale_colour_brewer(palette = "Dark2")+
    #scale_fill_brewer(palette = "Dark2")+
    scale_x_continuous(labels = scales::percent, limits=c(0,1))+
    labs(title=with(q_var_df, question[row_id==1]),
         subtitle = with(q_var_df, question[row_id==2]),           
         y=NULL, x=NULL) + 
    theme(axis.line.y = element_blank(),
          axis.ticks = element_blank(),
          plot.title = element_text(lineheight = 1.1, face="bold"),
          strip.text = element_text(size=14),
          legend.position = 'none')
}

plot_binary <- function(var_name, q_dfa=q_df, lbl_dfa=lbl_df, data_dfa=data_df){
 lbl_var_df <- lbl_dfa %>%
     filter(var==var_name, code==1) %>% 
  mutate(code=as.character(code))

 q_var_df <- q_dfa %>% 
  filter(var==var_name)

 clrs <- RColorBrewer::brewer.pal(7, "BrBG")[7]
 lbl_var_df$clr <- clrs
  
 df_plot <- data_dfa %>% 
    filter(var==var_name) %>% 
   semi_join(lbl_var_df, by=c("var", "code"))
 
 df_plot %>%
    filter(!is.na(code)) %>% 
    ggplot(mapping=aes(y=fct_rev(countryflag)))+
    geom_col(mapping=aes(x=pct, fill=(code))) +
    geom_text(data=lbl_var_df,
               aes(x=0, y=Inf, color=I(clr), label=label),
              family="Roboto Condensed",
               hjust="inward", vjust=1) +
    scale_fill_manual(values=clrs ) +
    theme_ipsum_rc(grid=FALSE)+
    scale_y_discrete(expand =  expansion(add = c(0,1)))+
    scale_x_continuous(labels = scales::percent, limits=c(0,1))+
    labs(title=wrap_txt(with(q_var_df, question[row_id==1]), 75), 
         y=NULL, x=NULL) + 
    theme(panel.grid.major = element_blank(),
          panel.grid.minor = element_blank(),
          axis.line.y = element_blank(),
          axis.ticks.y = element_blank(),
          axis.text.y = element_text(size=16),
          plot.title = element_text(lineheight = 1.1, face="bold", size=14),
          legend.position = 'none')
}


plot_binary_group <- function(var_group_name, q_dfa=q_df, lbl_dfa=lbl_df, data_dfa=data_df){

 q_var_df <- q_dfa %>% 
  filter(var_group==var_group_name, row_id==2) %>% 
  select(var, question)

 q_title <- q_dfa %>% 
  filter(var_group==var_group_name, row_id==0) %>% 
  pull(question)

 var_names_df <- distinct(q_var_df, var)
 
 lbl_var_df <- lbl_dfa %>%
     semi_join(var_names_df, by="var") %>% 
     group_by(code) %>% 
     summarise(label=first(label)) %>% 
     filter(code==1) %>% 
   mutate(code=as.character(code))

 clrs <- RColorBrewer::brewer.pal(7, "BrBG")[7]
 lbl_var_df$clr <- clrs

 df_plot <- data_dfa %>% 
    semi_join(var_names_df, by="var") %>% 
    left_join(q_var_df, by="var") %>% 
    semi_join(lbl_var_df, "code")

 df_plot %>%
    filter(!is.na(code)) %>% 
    mutate(question=map_chr(question, wrap_txt, 75)) %>% 
    ggplot(mapping=aes(y=fct_rev(countryflag)))+
    geom_col(mapping=aes(x=pct, fill=code)) +
    geom_text(data=lbl_var_df,
               aes(x=0, y=0.4, color=I(clr), label=label),
               family="Roboto Condensed",
               hjust="inward", vjust=1) +
    facet_wrap(vars(question), ncol = 2)+
    scale_fill_manual(values=clrs) +
    theme_ipsum_rc(grid=FALSE)+
    scale_y_discrete(expand =  expansion(add = c(1,0)))+
    scale_x_continuous(labels = scales::percent, limits=c(0,1),
                       expand = expansion(add=c(0.01, 0.02)))+
    labs(title=wrap_txt(q_title,105), 
         y=NULL, x=NULL) + 
    theme(panel.grid.minor = element_blank(),
          panel.grid.major = element_blank(),
          axis.line.y = element_blank(),
          axis.ticks = element_blank(),
          axis.text.y = element_text(size=20),
          plot.title = element_text(lineheight = 1.1, face="bold"),
          strip.text = element_text(size=12),
          legend.position = 'none')
}

When you click the Knit button a document will be generated that includes both content as well as the output of any embedded R code chunks within the document. You can embed an R code chunk like this:

Trust

COVID-19 Experience

COVID-19 effect

SARS

Prosocial

Canada Question

Finite Pool

Longitude

Personal

Friends

Media exposure and trust

Soughtinfo

Prep

# Multiselect plot

Govresponse

Govrestrict_1